The Cancer Gene Therapy Section's efforts focus on gene transfer approaches for the treatment of cancer. Gene therapy strategies for the treatment of cancer include the replacement of defective tumor suppressor genes in tumor cells, expression of antisense oligonucleotides against transforming oncogenes, expression of immune stimulatory molecules and cytokines, and the transfer of genes encoding enzymes that activate non-toxic drugs into cytotoxic agents in cancer cells. Despite initial optimism, clinical trials have reported few meaningful patient responses. This is believed to be a result of the low efficiency of in vivo gene transfer achieved by currently used vectors. A major effort of our laboratory has been the development of more effective viral vectors for gene transfer into tumors and efforts to enhance this limited transduction efficiency by vaccination. In an effort to enhance the gene transfer efficiency of adenoviral-based vectors, we generated a series of E1a-positive, E1b 55 kD-deleted and E1b 55 kD-enabled replicating adenoviral vectors expressing herpes simplex thymidine kinase (HSV-tk). These vectors are lytic and exhibit greater in vivo antitumor activity compared to a non-replicating E1-deleted HSV-tk-expressing vector. The E1b 55 kD-deleted vectors required administration of ganciclovir (GCV) for optimal tumor response, however, the E1b 55 kD-intact adenoviruses exhibited more robust replication and their efficacy was not improved by use of prodrug. Early (at 24 hrs.) GCV application inhibited the replication of these vectors. One of these vectors, Ad.OW34, an E1b 55 kD-intact vector caused significantly greater tumor regression and improved survival over a comparable E1b 55 kD-deleted or a non-replicating vector expressing HSV-tk in mouse models. The potential for dissemination and toxicity from replicating adenoviruses is an area of great concern for clinical application of these vectors. The safety testing of these vectors is hamper by the lack of a suitable animal model for replicating human adenoviruses. The cotton rat represents a semi-permissive model for replication of human adenoviruses, however, until recently there were no transplantable tumors in this animal model. In collaboration with Dr. Gregory Prince of Virion Systems, Inc. we are characterizing two unique tumor cells lines that support the replication of human adenoviruses and can be serially transplanted and form tumors in cotton rats. If confirmed, this would be an extremely useful model for the study and safety testing of these vectors. In a second approach, we engineered a novel retroviral vector for highly selective target cell gene expression using aspects of the natural life cycle of retroviruses. Moloney murine leukemia virus (MoMLV)-based vectors have been extensively used for clinical gene transfer owing to their ability to result in long-term stable gene expression in target cells. A major limitation, however, to the use of these vectors is the difficulty in generating stable retroviral vector producer cells (VPC's) capable of secreting viruses expressing toxic genes. Approaches used to overcome this problem have utilized tissue specific and inducible promoters, or complex recombination strategies. These various systems suffer from problems of inefficiency, or promoter "leakiness" leading to some level of gene expression in the VPC's, or activation of the transgene in the VPC's along with the target cells after treatment with the inducer molecule. To overcome this problem, we generated a series of MoMLV-based retroviral vectors encoding a REverse Transcription-ACtivated Transgene (RETRACT) that restricts gene expression to target cells. The prototypical RETRACT vector contains the cDNA of interest upstream of the viral 3' LTR in reverse orientation relative to the viral transcriptional unit. Without a promoter to drive the cDNA, no gene expression is seen. An exogenous promoter is cloned in reverse orientation at the R-U5 transition of the viral 5' LTR. On transduction of target cells by the RETRACT vector, the natural life-cycle reverse transcription of the retroviral RNA copies the 5'LTR with the promoter to the 3' LTR, where it drives expression of the reverse-oriented transgene. We studied this system using green fluorescent protein (GFP) as a marker gene and several different promoters including the SV40, CMV and RSV promoters and found no significant difference in gene expression. Detailed studies of the RETRACT system used the SV40 promoter. Southern blot and PCR analysis of genomic DNA from the VPC's and target cells demonstrated the expected approximation and orientation of the SV40 promoter to the GFP transgene only in the target cells. Northern analysis of the producer and the target cells demonstrated GFP transcripts of the appropriate length only in the target cells. Expression of GFP was not detected by flow cytometry in the PG13-RETRACT VPC's, however, fluorescence was seen in the target cells transduced with RETRACT VPC supernatants. To improve gene expression several variations of the RETRACT vector were generated including a single copy (sc), double copy (dc) and self-inactivating (SIN) versions of these vectors. The GFP fluorescence intensity was the greatest in a double copy vector (RETRACT/GFPdc). This work is currently under preparation as a manuscript. It is clear that most current gene therapy approaches to cancer represent local strategies that are limited by their inability to treat remote sites of disease or selectively target metastatic tumor. A novel vaccine approach currently in clinical trials is the use of peptide pulsed dendritic cells as a method of stimulating antitumor immunity. In this approach autologous dendritic cells incubated with synthetic peptide sequences containing mutations found in oncogenes or defective tumor suppressor genes expressed in the tumor are injected into patients in hopes of stimulating specific T cell responses against tumor. This strategy is predicated upon knowing the patient's HLA-type and that the binding affinity of a mutant peptide sequence to the MHC molecules are greater than that of the wild type protein sequence. An alternative approach involving transfer of the gene for the target antigen in to DC's offers potential advantages: (1) the tumor antigen can be expressed by the gene transfer vector without a detailed knowledge of the peptide epitopes or their binding affinities for a particular MHC molecule; (2) the DC's may naturally process the antigen leading to more effective presentation of epitopes and an improved immune response. Since joining the laboratory, Dr. Yoshio Sakai, has been working on adenoviral-mediated transfer of the genes encoding tumor antigens into dendritic cells as a cancer vaccine approach. The mutant gene encoding the oncoprotein or a non-functioning gene in the case of a dominant transforming oncogene is ex vivo transferred into autologous dendritic cells using high titer recombinant E1-deleted adenoviral vectors. The dendritic cells are then injected back into the subject where the expressed tumor antigen is processed and presented by the dendritic cells in the context of MHC to interact with T cells. Use of an adenoviral vectors may offer additional benefits over the synthetic peptide approach due to the adjuvant immunostimulatory effects of the native adenoviral proteins. We are studying this approach using HER-2/neu and the K-ras oncogenes as potential targets. Yoshio Sakai in the laboratory generated a series of recombinant adenoviral vectors encoding a truncated cDNA expressing the HER-2/neu extracellular and transmembrane domains (Ad.HER-2.ECDtm), ECD only (Ad.HER-2.ECD), or vector expressing no transgene (Ad.null). We studied effectiveness of adenoviruses to transfer reporter genes into cultured mouse bone marrow-derived dendritic cells. Despite published data to the contrary, we found that murine bone marrow-derived DC's could be efficiently transduced with an adenovirus expressing green fluorescent protein (Ad.GFP) at MOI's as low as 30 pfu of vector. Infection of immature DC's with adenoviruses induced maturation of the DC surface phenotype markers and co-culture of transduced DC's with splenocytes were able to stimulate greater proliferation in MLR cultures compared to uninfected DC's. We studied the effect of vaccination using dendritic cells transduced with these vectors in a transgenic mouse model of HER-2/neu induced breast cancer. BALB/c-neu T mice transgenic for the rat HER-2/neu oncogene under the control of an MMTV promoter spontaneously develop breast cancers beginning at 15 weeks of age. Five-to-seven week old BALB/c-neu T mice were injected weekly x 3 weeks with 106 DC's treated with Ad.HER-2.ECDtm, Ad.HER-2.ECD, or Ad.null. Another group received unmodified DC's alone. At 26 weeks, all of the mice treated with DC's transduced with Ad.HER-2.ECDtm and 20% of mice treated with DC's transduced with Ad.HER-2.ECD were free of tumor. In contrast, all of the mice (100%) treated with the DC's infected with Ad.null or with unmodified DC's have developed breast cancers. Studies are underway to characterize the type of immunity generated this gene therapy-dendritic cell vaccine. In collaboration with the laboratory of Dr. Jay Berzofsky, we are examining this approach in a second model, we generated murine MC38 colorectal carcinoma cells as well as murine sarcoma cells generated from an A2-transgenic mouse (FA2Kb) expressing mutant human K-ras genes (G12V or G12D) and recombinant adenoviruses expressing non-signaling forms of these two mutant K-ras genes as well as epitope enhanced versions of these vectors. Preliminary studies show that these modified cell lines express the mutant gene and form palpable similar to the wild type MC38 cells in C57BL/6. Animal studies are in currently in progress or planned using these models.